This table should be used as a power configuration guide.  Powers shown are approximate and probably maximum depending on the size and quality of the power supply.

From Post #283 by Quasi 20th December 2005 05:31:15:

The following table can be used as a guide when trying to determine output configurations for the amp module.  It provides recommendations for FETs per amp module (not FETs per rail).

I have based the table on typical IRFPXXX FETs rated at 180 watts (Tc = 25 degC), derated for a Tc of 50 degC.  Naturally more / less FETs can be used if they are weaker / stronger.

The table does assume a virtual limitless power supply so actual power acheivements will probably be less.

The table does not remove the need to fully understand the workings of the 3 other main parts of the schematic as there are some voltages that need to maintained irrespective of the supply rails.

From Post #1457 by Keypunch 26th January 2007 18:51:27:

I have updated the Table of Post # 285 to include the R6 value based on Vrail and fusing for the required load power in my comments of Post #1190.

Please note that the above referenced table has been reposted as is a few times in the thread.  I do not have the time to find those postings to list them in reference to the updated table attached with the fusing and R6 columns added.

I have used Andrew's suggested method in Post #1451 to calculate fusing.

Please note that this table does not agree with quasi's recommended fusing shown on the quasi PCB layouts.  Quasi's recommended fusing for 6 FETs for 4 ohms is 5 Amps, for the 10 FETs for 4 ohms is 10 Amps and the 4 TO-220s/2 TO-247s based design of 4 Amps which I believe is fusing rated for 4 ohms.  I am not sure if quasi's fuse recommendations are based on using slow blow fuses, which may account for the lower fuse rating quasi suggests.

From Post #1897 by Keypunch 22nd June 2007 12:29:13 with typo correction:

I have taken the time to update your "Power selection guide" on the NMOS350 / 500 of your site.  I have added a specific reference to the number of FETs as being IRFP450 to avoid any confusion of those that wish to use a different MOSFET output device.  The R6 is from your latest guidance of R6.  Although I know fusing may depend on various factors, your schematics do include fusing recomendations so I have used the guidance of the thread discussion to determine the various fusing values.

There are some other R6 values proposed in the past. In one case a builder had an interesting experience with the rail voltage he wished to use, but using different rail voltages above and below the rail voltage he wished to use were ok.  The builder changed a few reistor values, including R6.  The details of the previously suggested R6 values and the R6 value the one builder used are indicated below for reference in case you wish or need to evaluate alternative R6 values.

From Post #159 by quasi 18th March 2005 22:29:37
(at time the same R6 Resistor was known as R7 for this much older schematic):

2. For lower voltage rails the main change that has to be made to my circuit is resistor R7 in the first stage.  This shares the negative rail voltage with T2 and must leave sufficient voltage for T2 to work properly.  The constant current source is setup to provide about 2.7 mA and this causes a voltage drop across R7 of 45 volts.  With full power rails of 57 volts approximately 12 volts is left across T2 so everything is ok.  I would make sure that T2 always had at least 5 volts across it to allow for power fluctuations.  For rails lower than 50v I would change R7 to 10k.

From Post #292 by quasi 20th December 2005 17:40:58:

Also with 45v supply you need to change R6 to 12K otherwise the constant current source (T4) will not work properly.

From Post #325 by quasi 17th January 2006 03:53:07:

Example 3: The approx voltage across
T4 = negative rail voltage - v(R7) - v(R6).  With 70 volt rails this works out to be 70 - 0.6 - 46.8 = 22.6v.  By changing R6 the voltage across T4 can be varied.

From Post #368 by quasi 8th February 2006 02:33:46:

To run off +/- 40v resistor R6 must be changed from 18K to 10K or can be left out and a link used instead.

From Post #438 by quasi 13th April 2006 05:48:33:

You can use any low signal / low noise transistor for the constant current source.  Watch the Vce though (rail - V R6).

From Post #1025 by quasi 2nd September 2006 21:01:42:

R6 needs to be changed for different rails.

> 60 volts R6 = 18K
30 - 60 volts R6 = 6K8
less than 30 volts R6 = 0 ohms (link)
This must be done or the first stage won't work properly.

From Post #1827 by Samuel Jayaraj 29th May 2007 07:11:32:

Well, after months of infrequent revisits to my implementation of the Quasi N-channel amp, I was able to get the amp to work correctly.

Let me describe the problem I faced:  First, the PCB was my own design, more like a conventional design with all the power Mosfets aligned to one side of the board.  I used 3 x IRFP450 per half, all other parts were standard and as per schematic, the only exceptions being that the MJE340/350 parts were KSE340/350 plastic insulated types.  Voltage was +-63 volts DC.  I used a different circuit for both soft start and DC protection.  Leach's clip indicator was also made integral on the amp board.  (Input transistors are very much 2SC1845s.)

The problem was that the music had a crackling noise/distortion with every beat even at very low levels.  I tried Iq from 100 mA to 300 mA but the phenomenon did not go.  Initially this phenomenon occured only when playing both channels.  I thought this might be a ground issue and rewired the amp after which the distortion was heard even while playing one channel only.

However, Iq and offset could be set very precisely and there was absolutely no drift.

I tried other voltages viz., +-35vDC, +-42vDC, +-56vDC apart from +-63vDC.  I did not notice the problem with +-35 and +-42 volts.  Hence, I changed R6 to 22K and R12/R15 to 68E, in place of 18K and 47E respectively.  I checked only one channel with +-56 volts DC this morning and the music was clean.  Sounded pretty good too.

Only could not bring down offset below 63mV.  I need to check whether R6 or R12/R15 has effected the change.

My question is whether there is any difference between MJE and KSE (plastic insulated version) types?

Has anyone else run Quasi's amp with high rail voltages and standard parts without any problems?

What could be possible reasons for the phenomenon I encountered?

This is only a preliminary report and I will give more details hopefully tomorrow, but this is something I thought I'd share for others' insights and comments.

Thanks,

__________________
Sam

From Post #1837 by Samuel Jayaraj 30th May 2007 05:48:53:

Yesterday I had posted regarding some of the issues I am facing with Quasi's Mosfet Amp.

Today I tried again and I was able to set offset to zero volts.  With rails of +-56 volts, Iq of 110mA and 3 pairs of IRFP450s the sound is not as good as playing the same amp with rail voltages of +-35volts or +-42 volts, but with R6 and R12/R15 in original values as per schematic.

The changes I have made are changing R6 to 22k and R12/R15 to 68E.  With original values and rail voltages upwards of +-56 there is audible distortion.

__________________
Sam

From Post #1841 by Samuel Jayaraj 31st May 2007 08:26:53:

This is how I configured the final setup of my Quasi Amp:

Now the amp is really upto the challenge.  Rhythm and pace are among the very best.  Very good resolution of low level signals.  Sound stage is very controlled and not aggressive or harsh.  Low Bass and Mid Bass are very, very good.

Does not image as forward as my AKSA clone, Stochinno Amp and Gain Clone (LM3886TF), in that order; however, the mids are not withdrawn or shy.  I would have liked the overall image to be a little more forward, personally speaking.  But the musicality and timing more than make up for this.

T6/T7 do run a bit hot - have to increase the heatsinking for these.

Any comments or opinions?

E and R both represent Ohms.

__________________
Sam

From Post #1846 by Samuel Jayaraj 31st May 2007 11:51:33:

Quasi, you are right, that was a typo. R12/R15 are indeed 68 ohms.

I did try with and without splitting the rails but I find that isolating the Front End almost always cleans up the Bass and Mid/Bass resulting in a more resolving Mid band; it was the same case with this amp.  However, I should say that with the original schematic values, the distortion that I am talking about did slightly reduce when the supplies were joined.  The value that I used initially for the resistor was 100 ohms, following the diode and the caps were 100uF/0.1uF.  This did increase the distortion as compared to joining the rails, but in both cases, the distortion did not disappear.

T6/T7 are on a heatsink but they run quite hot, NOT warm.  Right now the amp is out of its case and the temperature is bound to rise in a closed (ventillated) box.

Resistors with values in the range 0.05 to 0.10 ohms with capacitance banks more than 10,000uF on either side help to clean up supply hash below 500Hz!  So did Elektor publish in their 60 watt HEXFRED amp article.  I have never measured this, but I have invariably heard improved Bass with this arrangement.

After changing R12/R15 to 68 ohms and varying R6 from 6K8 to 18K, the distortion was almost gone.  But the sound was dull, lifeless and did continue to distort on music peaks.  Hence, I changed R4/R8 to 390E to begin with and then further reduced them to 330E.  With this arrangement and R6 at 18K, the Mids were a little more recessed and the amp was not quite up to it in speed.  Overall musicality was lacking.  When I changed R6 to 10K, the speed/pace, rhythm and musicality were really good.  ( I should add that from output to the junction of R18/C7, I have a 47K resistor.  The brings down noise and improves sonics. It does change the feedback factor but whenever possible, I try to use this technique.  A very famous DIY amp used this technique as well.)

Andrew, I did increase Iq right upto 300mA since I have quite big heatsinks.  But this was prior to solving the distortion problem and having stock values of components.  In the current setup, I have not tried anything more than 110mA.

I guess increasing C7 will result in less low frequency phase shift, but I have not tried it because I am using Elna RJH 100uF caps, in shunt with 0.1uF and 2 x IN4148 anti-parallel diodes. I like the sound of the Elnas. I could probably try two in parallel.

For C1, I do have 1uF, 2.2uF, 2.7uF, 3.3uF in high quality film types and also 4.7uF Solen MKP capacitors.  I could have tried them but have not yet.

What is the effect of increasing the value of C14?

I am not pushing the amp into high drive levels, because my speakers are 91.5db sensitivity in a Quarter Wave Loaded cab and hence, are quite efficient.  I also have a pair of Fostex 168 Sigma ES, but have not used them in these cabs, in which they fit and play well.  The speakers are 4 ohms, 2-way and I have been alternating between 4 ohms and 2 ohms while testing the amp, just to be sure that at higher current delivery, the behaviour of the amp doesn't change.  At the drive levels required for my speakers, the 2 ohm performance of the amp is audibly the same as its 4 ohm performance.

In the next round I'd also like to split R11 and put a cap from this junction to the negative rail.

Quasi and Andrew, thanks for your respones. Do look forward to more inputs and discussions.

__________________
Sam

Last edited by Samuel Jayaraj on 31st May 2007 at 11:57:19

From Post #1451 by AndrewT 23rd January 2007 11:03:37:

Hi,
the rail fuses are there for medium to long term protection, just as the heatsink is starting to get hot.

Prior to that, the electronic current limit is operating to just keep the output devices INSIDE their SOAR but the disspation at the limiting current causes rapid overheating. Continued heating will bring the SOAR down to meet and then pass the limiting current value. Then the output stage blows.

The chosen fuse must blow before the Tc rises above the acceptable temperature chosen for the current limit and SOA.

as an example;
limiting current about 8A.
Vrail +-60V each transistor will dissipate 480W for a half (square wave)
cycle. the heatsink will heat very rapidly.
Fuse blows within 200mS to 1S.

I use a fast fuse rated at half peak current into minimum rated load.

Vrail=60V and 4ohm speakers Ipk~=56/4=14Apk, fuse =F7A and due to time delay in a fast fuse they never cause nuisance blowing and very loud music. Fusing this way even allows full power testing into rated load (due to halfwave rms value).

__________________
regards Andrew T.